Workshop: Biosystematics by Julian Lee (revised by D. Krempels) Biosystematics (sometimes called simply "systematics") is that biological sub-discipline that is concerned with the theory and practice of classifying organisms. The distinction is sometimes made between taxonomy, which is concerned with actually naming and classifying organisms, and systematics, a more inclusive term that subsumes taxonomy, but also includes the philosophical, theoretical, and methodological approaches to classification. In practice, the terms are often used interchangeably. Most biologists agree that our classifications should be "natural" so that to the extent possible they reflect evolutionary relationships. We do not, for example, place slime molds and whales in the same family. Biosystematics, then, is a two-part endeavor. First, one must erect an hypothesis of evolutionary relationship among the organisms under study. Second, one must devise a classificatory scheme that faithfully reflects the hypothesized relationship. We will use a series of imaginary animals to introduce you to two rather different methods that attempt to do this. The hypothetical animals used in this exercise are called Caminalcules and were created and "evolved" by J. H. Camin at the University of Kansas in the 1960's. They have served as test material for a number of experiments concerning systematics, its theory and practice. Use of imaginary organisms for such studies offers a distinct advantage over using real groups, because preconceived notions and biases can be eliminated. Phenetics (Numerical Taxonomy) Those who use this approach to inferring evolutionary relationships group organisms on the basis of their overall similarity. In this portion of the workshop, you will become familiar with a common numerical taxonomic method (Cluster Analysis) by working an actual problem involving selected Caminalcules (Figure 1). In so doing, you will come to understand the important assumptions of this methodological approach, its strengths and its weaknesses. Figure 1. A variety of Caminalcules, arranged in no particular order.
A Sample Procedure Carefully examine the eight Caminalcules illustrated in Figure 1. These will be your Operational Taxonomic Units (OTUs)--a name we use to avoid assigning them to any particular taxonomic rank (such as species). You may think of them as biological species, and refer to them by number. The first step is to make a subjective judgement about the overall similarity between all pair-wise combinations of the eight OTUs, using a scale of 1.0 (maximum similarity) to 0 (complete disimilarity). These similarity rankings are then cast into a similarity matrix. An example of such a matrix (for the 8 OTUs in Figure 1) is shown in Table 1. Table 1. An example of similarity rankings among the eight Caminalcules pictured in Figure 1. The rankings have been subjectively assigned. 1 2 3 4 5 6 7 8 1-2 0.1-3 0.2 0.1-4 0.7 0.3 0.4-5 0.5 0.2 0.8 0.3-6 0.8 0.2 0.4 0.7 0.4-7 0.1 0.9* 0.2 0.3 0.2 0.3-8 0.5 0.3 0.6 0.4 0.6 0.4 0.4 - Step 1. Find the pair of OTUs that have the highest similarity ranking. (In this example, it happens to be OTUs 2 and 7, with a similarity ranking of 0.9 shown in boldface and with an asterisk*). Step 2. Combine OTUs 2 and 7, and treat them as a single composite unit from this point on. Construct a new similarity matrix (this time it will be 7 x 7), as shown in Table 2. Table 2. Reduced matrix with similarity values recomputed for all OTUs with composite OTU 2/7. 2/7 1 3 4 5 6 8 2/7-1 0.1-3 0.15 0.2-4 0.3 0.7 0.4-5 0.2 0.5 0.8* 0.3-6 0.25 0.8* 0.4 0.7 0.4-8 0.35 0.5 0.6 0.4 0.6 0.4 - Step 3. Recalculate the similarity values for each OTU with the new composite 2/7 OTU. To do so, simply compute the average similarity of each OTU with 2 and with 7 (listed in Table 1). In our example, the similarity of OTU 4 with OTU 2 is 0.3, and the similarity of OTU 4 with OTU 7 is 0.3, then the similarity of OTU 4 with OTU 2/7 will be the average of those two similarities: [(0.3 + 0.3)/2], or 0.3. The recalculated similarities for each OTU with composite OTU 2/7 are shown in boldface in Table 2. The new highest similarity value(s) is/are marked with an asterisk*. Step 4. In the new, reduced matrix with recomputed similarity values, find the next pair of OTUs with the highest similarity value. In this case, OTUs 1 and 6 and OTUs 3 and 5 are tied with a similarity value of 0.8. For simplicity, choose one pairing at random and recalculate the similarity indices, and then do the next pairing, as shown in Table 3(a).
Table 3. Reduced matrices with similarity values recomputed for all OTUs with (a) new composite OTU 1/6, (b) new composite OTU 3/5, (c) new composite OTU 4/1/6, and (d) new composite OTU 8/3/5 and (e) 4/1/6/8/3/5. Again, recalculated similarity indices are shown in boldface, and the new highest similarity index in each matrix is marked with an asterisk. a: 1/6 2/7 3 4 5 8 1/6-2/7 0.18-3 0.3 0.15-4 0.7 0.3 0.4-5 0.45 0.2 0.8* 0.3-8 0.45 0.35 0.6 0.4 0.6 - b: 3/5 1/6 2/7 4 8 3/5-1/6 0.38-2/7 0.35 0.18-4 0.35 0.7* 0.3-8 0.6 0.45 0.35 0.4 - c: 4/1/6 3/5 2/7 8 4/1/6-3/5 0.37-2/7 0.24 0.35-8 0.43 0.6* 0.35 - d: 8/3/5 4/1/6 2/7 8/3/5-4/1/6 0.4* - 2/7 0.35 0.24 - e: 4/1/6/8/3/5 2/7 4/1/6/8/3/5-2/7 0.3* - Step 5. Continue to construct reduced matrices, each time recalculating the similarity indices between your new composite OTU with all remaining OTUs, as shown in Table 3(b - e). The last step will result in a 2 x 2 matrix with a single, final similarity value. (In this example, composite OTY 4/1/6/8/3/5 has a 0.3 similarity to composite OTU 2/7). Step 7. Your OTUs can now be clustered graphically in a branching diagram called a phenogram. The result of our sample cluster analysis is shown in Figure 2. Note that the most similar OTUs from the first table (OTUs 2 and 7) have been paired at a branch point reflecting their similarity index (0.9). The next most similar OTU pairings (1/6 and 3/5) have been clustered at their respective levels (0.8) in the same fashion, and each successive reduced matrix yields the appropriate branch level, shown on the similarity scale at the left side of the diagram. This phenogram is designed to not only indicate which of the Caminalcules are supposedly most physically similar, but also to what degree they are phenotypically similar.
Figure 2. Phenogram of Caminalcules clustered in the sample exercise. Note again that the similarity values assigned in the example were subjective. Different relative similarities might have been assigned by a different observer. A phenogram provides a visual representation of similarity relationships of OTUs, and nothing more. Note that clusters can be rotated around their branches without changing the implied similarity relationships. For example, if the diagram were rotated at the 0.3 branch point so that OTUs 2 and 7 were on the right, and OTUs 1, 3, 4, 5, 6 and 8 on the left, the information in the diagram would not be changed. The same would be true of rotation at any branch point. Given the computational tedium of working through this simple 8 OTU example, you can see why real numerical taxonomic studies, which often involve large numbers of OTUs, require the use of computers.
Exercise You will now do your own phenogram using a somewhat different group of Caminalcules as your OTUs. A series of tables is provided below for you to do your reduction calculations, and the space on the following page is provided so that you may draw a phenogram for your Caminalcules, shown in Figure 3. Once you have finished the exercise, discuss the questions following the exercise. Figure 3. A second set of Caminalcules, arranged in no particular order for you to use in your own phenetic analysis (and later, in a cladistic analysis). Questions 1. In the early years of numerical taxonomy, advocates of teh approach claimed that the methodology would bring objectivity to systematics by the appliation of highly quantitative techniques, specific algorithms for clustering, and numerical assessment of similarity. What to make of these claims of objectivity? 2. An underlying assumption of numerical taxonomy is that overall similarity is a reliable indicator of evolutionary relationships (i.e., recency of common ancestry) among organisms. Is this a reasonable assumption? 3. If numerical taxonomy is such a flawed method for inferring evolutoinary relationships, why are we torturing you with all this? (Actually, this is a serious question: Do you believe that cluster analysis might be of use in other areas of biology? If so, in what fields, and how?)
Matrices for you to use: A B C D E F G H A - B - C - D - E - F - G - H -
Cladistics (Quantitative Phyletics) A more objective and suitable method for inferring evoluitionary relationships is cladistics, sometimes called quantitative phyletics. Rather than grouping OTUs together on the basis of overall similarity, as in numerical taxonomy, the investigator using this method groups OTUs together on the basis of shared, derived characters--characters whose presence or absence in two or more OTUs is inferred to be the result of inheritance form their common ancestor. Thus, in the branching diagram shown in Figure 4, OTUs A and B are considered most closely related if they share characters in common not found in OTU C. OTU C is considered most closely related to A and B if it shares characters common to all three OTUs, but absent from other OTUs outside this three-otu grouping. Results of a cladisitc analysis are normally summarized as a branching diagram called a cladogram (from the Greek clad meaning "branch") (Figure 4), which is an explicit hypothesis of evolutionary relationships. In the cladogram shown, A and B are considered most closely related if they share characters in common that are not found in C. C is considered most closely related to A and B if it shared charactrers common to all three OTUs, but absent from other OTUs outside this three-otu group. The cladogram is an explicit hypothesis of evolutionary relationship. A taxonomic group descended from a single common ancestor is said to be monophyletic if it includes all the descendants of that common ancestor, and none of the taxa descended from a different common ancestor. For example, A and B would comprise a monophyletic group, but A and C would not. Figure 4. A cladogram A Sample Procedure We will examine the eight Caminalcules in Figure 1, this time in an attempt to infer the evolutionary relationships among them, as follows. Step One. Select a series of characters that can be expressed as binary (i.e., two-state). For example: Character a: "eyes present" (+) versus "eyes absent" (-) Character b: "body mantle present" (+) versus "body mantle absent" (-) Character c: "paired, anterior tentacles present" (+) versus "paired, anterior tentacles not present" (-) Character d: "anterior tentacles flipperlike" (+) versus "anterior tentacles not flipperlike" (-) Character e: "eyes stalked" (+) versus "eyes not stalked" (-) Character f: "body mantle posterior bulbous" (+) versus "body mantle posterior not bulbous" (-) Character g: "eyes fused into one" (+) versus "eyes separate" (-) Character h. "forelimbs with digits" (+) versus "forelimbs without digits" (-) Step Two. Examine all your organisms and determine which character state it exhibits. Enter the data in a matrix like the one shown in Table 4.
Table 4. Character states of characters a - h in Caminalcules in Figure 1. character 1 2 3 4 5 6 7 8 a + + + + + + + + b + + + + + + + + c - + - - - - + - d + + - + - + + + e - + - - - - + - f + - - + - + - - g + - - - - - - - h - - + - + - - - Note that in this example, character a (presence or absence of eyes) and character b (presence or absence of a body mantle) is the same in all eight OTUs. Hence, this (primitive) character is not useful to us if we are trying to determine differences between the OTUs. (A similar example would be "hair" in mammals. Since all mammals have hair, the presence of hair in any particular mammal gives no information about its relationship to other mammals.) Note also that only OTUs 2 and 7 share character e (stalked eyes), which is absent from all other OTUs. This suggests that OTUs 2 and 7 both inherited this character from a common ancestor. Likewise, OTUs 1, 4, and 6 share character f (bulbous mantle posterior) which is absent from all others. This supports the hypothesis of common ancestry among these three OTUs.. The same reasoning argues for common ancestry among OTUs 1, 4, and 6 (character f), and for OTUs 3 and 5 (character h), and so on. A cladogram consistent with the distribution of these eight characters among the eight OTUs is shown in Figure 5. Figure 5. A cladogram based on shared, derived characters in Caminalcules 1-8.
It is important to note that this is not the only possible branching sequence consistent with the distribution of the characters among the OTUs. In practice, there are often several, or even many, cladograms that can be constructed, all of which are consistent with the data. In such cases, systematists generally apply a parsimony citerion for selecting the "best" cladogram. The rule of parsimony states that when two or more competing hypotheses are equally consistent with the data, we provisionally accpet the simplest hypothesis. This is not to say that evolution is always parsimonious, only that our hypotheses should be. In the case of competing cladograms, the rule of parsimony would require that we accept the simplest cladogram--the one with the fewest "steps" to each of the taxa on the tree--is preferred over a cladogram requiring more evolutionary steps. In our example, we could hypothesize that OTU 6 is actually more closely related to OTU 1 than to OTU 4. However, this would require that character g (fused eyes) had been evolved once, and then secondarily lost in both OTUs 4 and 6. This is a less parsimonious hypothesis than one stating fused eyes evolved only once, in OTU 1. Classification Given an hypothesis of evolutionary relationships, the second step in biosystematic endeavor is to erect a classification that faithfully reflects those relationships. Because the results of a cladistic analysis (i.e., the cladogram) are heirarchical, they can easily be incorporated into the Linnaean hierarchy, as follows. Order Caminalcula: Family 1 Genus 1 Species 2 Species 7 Genus 2 Species 1 Species 4 Species 6 Species 8 Family 2 Genus 1 Species 3 Species 5 In cladistic analysis, all taxa must be monophyletic, meaning that they must include the common ancestor (almost always hypothetical) and all descendants of that common ancestor. Thus, in the cladogram above, OTUs 2 and 7 together with their common ancestor (at the branch point just below them) constitute a monophyletic genus, as do OTUs 1,4,6 and 8 and their common ancestor (at the branch point just above the appearance of character d). A Family consisting of only OTUs 2 and 7 would not be monophyletic, because it does not include all the descendants of the common ancestor (at the branch point just below character d). Such a group would be considered paraphyletic (containing some, but not all, of a particular ancestor's descendants). A Family consisting of OTUs 2 and 7 plus OTUs 3 and 5 would be considered polyphyletic (consisting of species derived from more than one most recent common ancestor). This is because such a taxon would be made up of groups descended from both the ancestor just below the appearance of character h, and the one just below the appearance of characters c and e. Exercise Using the Caminalcules in Figure 3, go through the steps of cladistic analysis we just did for the ones in Figure 1. Use the spaces and matrix provided to choose shared, derived characters that help you link some of the OTUs in taxa that reflect their hypothetical evolutionary relationships. Finally, in the space provided, draw a cladogram of your Caminalcules, showing the appearance of each character as we did in our example.
character state of character if (+) state of character if (-) a b c d e f g h OTUs character 1 2 3 4 5 6 7 8 a b c d e f g h Your cladogram: Questions 1. The results of phenetic and cladistic analysies are inherently hierarchical, as is the branching sequence of the evolution of organisms. So, too, is the Linnaean classificatory system of ever more inclusive taxonomic categories from species, Genus, Family, Order, Class, Phylum, Kingdom and Domain. Can you name some other hierarchical classification systems (not necessarily biological)? 2. Taxonomists were erecting classifications of organisms long before Darwin convinced biologists of the reality of evolution. Some of these taxonomists believed in the fixity of species and in special creation. Nevertheless, in some respects, these pre-darwinian classifications are rather similar to those produced later by evolutionary taxonomists. Why do you think this is so?